Synthesis and antimicrobial activity of a carbocyclic puromycin analog

Chem. , 1972, 15 (2), pp 171–177. DOI: 10.1021/jm00272a012. Publication Date: February 1972. ACS Legacy Archive. Cite this:J. Med. Chem. 15, 2, 171-...
0 downloads 0 Views 962KB Size
Journal ofMedicina1 Chemistry, 1972, Vol. 15, No. 2

arbocyclic h r o m y c i n Analog

concns of inhibitor. The in vitro antitumor assays were carried out by our microassay technique which involves the introduction of 0.5 ml aliquots of the medium (RPMI 1630 + 10% calf serum) contg the various concns of the analog into 16 X 125 mm screw cap culture tubes, followed by 0.5 ml portions of medium contg 3 X lo5 L1210 cells. The cultures are incubated at 37" for 40 hr, after which the viable cells are counted by trypan blue exclusion. During this time the cell number in the controls increases approximately eight- to ninefold, with an average viability of 99%.

Acknowledgment. The excellent technical assistance of Miss Ginger Dutschman and Mr. R. J. Maue is gratefully acknowledged. This study was aided by grants CA-12585 and 12422 from the U. S.Public Health Service and T 4 3 6 from the American Cancer Society. References (1) H. Nishimura, K. Katagiri, S. Kozaburo, M. Mayama, and N. Shinaoka, J. Antibiotics, Ser. A , 9, 60-62 (1956).

171

(2) K. Ohkuma, ibid., Ser. A , 14,343 (1961). ( 3 ) M. Sameyoshi, R. Tokuzen, and F. Fukuoka, G a m , 56,219 (1965). (4) W. L.Wilson, Cancer Chemother. Rep., 52,301 Illus. (1968). (5) E. C. Taylor and R. W. Hendess, J. Amer. Chem. Soc., 87, 1995 (1965). (6) C. W.Noell and R. K. Robins, J. Heterocycl. Chem., 1,34 (1964). (7) R. L.Tolman, R. K. Robins, and L. B. Townsend, J. Amer. Chem. Soc., 91,2102(1969). (8) R. L.Tolman, G. L. Tolman, R. K. Robins, and L. B. Townsend,J. Heterocycl. Chem., 7,799 (1970). (9) M. Bobek, R. L. Whistler, and A. Bloch, J. Med. Chem., 13, 411 (1970). (10)E. J. Reist, D. E. Gueffroy, and L. Goodman, J. Amer. Chem. Soc., 86,5658 (1964). (11) A. Bloch, E.Mihich, C. A. Nichol, R. K. Robins, and R. L. Whistler, Proc. Amer. Ass. Cancer Res., 7,I (1966). (12) B. Urbas and R. L. Whistler, J. Org. Chem., 31,813 (1966).

Synthesis and Antimicrobial Activity of a Carbocyclic Puromycin Analog. 6-Dimethylamino-9- { R- [ 2R -hydroxy-3R -( p-methox yphenyl-r,-alan ylamino ) ]cyclopentyl)purine? Susan Daluge and Robert Vince* Department o f Medicinal Chemistry, College of Pharmacy, University of Minnesota, Minneapolis, Minnesota 55455. Received July 23, I971

An assessment of the requirement for the furanosyl 0 and the CH'OH moiety in the puromycin molecule was undertaken by the synthesis of a novel puromycin analog. A carbocyclic analog, 6-dimethylamino9-(R-[2R-hydroxy-3R-(p-methoxyphenyl-~-alanylamino)lcyclopentyl )purine (2), was synthesized and evaluated for antimicrobial activity. The carbocyclic analog exhibited antimicrobial activity comparable to puromycin, and also circumvented the nephrotic syndrome associated with puromycin by releasing a nontoxic aminonucleoside upon hydrolysis. The diastereoisomer (19) of 2 was also isolated and found to be devoid of antimicrobial activity. Puromycin (l), an antibiotic with antitumor activity,' has been found to inhibit protein synthesis in a wide variety of organisms. Its structure has a striking resemblance to that of the aminoacyl-adenyl terminus of aminoacyl-tRNA, and it has been demonstrated that the antibiotic causes premature release of the polypeptide chains from the ribosome.' For this reason, puromycin has been used extensively as a tool in the investigation of protein biosynthesis.

1

A variety of analogs and isomers of puromycin have been prepared to define the structural requirements for inhibition in an attempt to further understand its mode of a c t i o n . j ~ ~ ?This work w a s generously supported by Grant AI 08142 from the

US.Public Health Service.

However, all of these structures have been of the classical nucleoside type in which an N-substituted amino sugar is attached to a purine or pyrimidine ring through a glycosidic linkage.36 The difficulties encountered in preparing 3aminoribosyl nucleosides have severely limited the availability of these compounds. Also, the classical nucleoside compounds introduce two undesirable structural features into the puromycin analogs which have not been demonstrated as essential for biological activity; Le., the furanosylo and the 5'-OH group. Thus, it may be possible to modify 1 within the region outlined by the dotted line and still retain the activity of the antibiotic. Since ribonucleosides are easily cleaved hydrolytically or enzymatically, many nucleosides which may be effective chemotherapeutic agents become ineffective in vivo because they are rapidly destroyed by cleavage into a purine or pyrimidine and a carbohydrate moiety.''* This difficulty could be circumvented by replacing the furanosyl ring with a cyclopentyl system which sterically simulates the sugar moiety and provides a hydrolytically stable C-N bond. The removal of the 5'-OH group from puromycin and its analogs would be desirable from a toxicity standpoint. Toxic manifestations, including renal lesions, have precluded the use of puromycin in the treatment of human or animal infectious diseases or neoplasm^.^ The nephrotic syndrome results from small amounts of aminonucleoside produced by the hydrolytic removal of the amino acid moiety from administered p u r ~ m y c i n .Recent ~ studies demonstrate that the aminonucleoside is first monodemethylated" and subsequently converted to the 5'-nucleotide." It has been s u g

172

Journal of Medicinal Chemistry, I 9 72, Vol. 15, No, 2

Daluge and Vince

Scheme I

Et,O-H,O

0

dioxane

4

3

EtOH

5

7

6

NMe,

NMe, I

I

1 . NH,OH'HCl

0

HON

9

8

10

1 . B,H,

12

THF

Ac,O-p yridine

14,R= H

13, R = H 13a, R = Ac

14a, R = Ac

CbzCl Ba(0H) l 3 Tq++

DM F

o

HN I

cbz

n

DMF

W

HN)(o 0

17

18

:->

d 6 HN

15

co

co RHNfH

RHNfH

OMe 19, R = n 19a, R = Cbz

OH

I

I

OMe

z

R=H Z;, R = cbz

gested that this 5'-nucleotide is responsible for the nephrotic syndrome. In view of these observations, we decided to synthesize the carbocyclic analog 2 which retains the structural reauirements that have thus far been demonstrated to be esskntial for activity., The requirement for the 2'-OH in puro-

11

OH

AcON

OAc

12

mycin has not been demonstrated. However, it is required for the acceptor activity of aminoacyl adenosines12 which are thought to bind at the ribosomal site in the same manner as puromycin. Thus, the antimicrobial activity of the carbocyclic analog, 2, would provide information relating to the structural requirements of the 4'and 5' oxygens of puromycin for biological activity. In addition, 2 represents a new class of puromycin analogs which should produce a nontoxic aminonucleoside upon hydrolysis of the amino acid moiety. Chemistry. The carbocyclic analog (2) of puromycin was synthesized by the route shown in Scheme I. (Structures 318 depict only one enantiomer of the racemic form actually obtained.) Cyclopent-2-enone ethylene ketal (3)13 was treated with NBS in Et,O-H,O by a modification of the procedure of Guss and RosenthalI4 in which NaHC03 was used to buffer the reaction mixture. The resulting bromoh drin 4 decomposed on storage and was not characterized.rWhen bromohydrin 4 was immediately treated with NaOH in refluxing benzene, epoxide 5 was formed in 73% yield (from 3), as a stable colorless liquid. The epoxide was opened with NaN, in dioxane by the method of Vander Werf, et al.,l5 and the resulting azide reduced catalytically to crystalline trans-3-amino-2-hydroxycyclopentanone ethylene ketal (7) in 76% yield (from 5). Attack by N3- at C-3 of epoxide 5 would be expected.I6 The structure of 7 was confirmed by the nmr spectrum in which H-2 appears as a doublet at T 6.53 (J = 7.9 Hz), and H-3 appears upfield as a multiplet at T 7.36.8. The purine moiety was formed via a standard method.17 Condensation of 7 with 5-amino-4,6-dichloropyrimidine, followed by ringclosure of the resulting crude pyrimidine 8 with triethyl orthoformate in the presence of EtS03H gave the 6-chloropurine 9 in 84% yield (from 7). The 6-dimethylaminopurine 10 was formed in 75% yield when 9 was treated with refluxing Me2NJ3. The ketal function of 10 proved unusually resistant to hydrolysis. Attempted hydrolysis by refluxing HC1 at pH 3 , refluxing aq NH4Cl, or transketalization in acetone with pTsOH gave only recovered 10. This difficulty has been encountered in the hydrolysis of other ketals and acetals of similar purine derivatives.§ The proximity of a protonated $The corresponding bromohydrin of cyclohex-2-enone ethyleneketal prepd b y this method was a stable, cryst solid, mp 96-98', which was characterized; unpublished results.

8 Unpublished

results.

Carbocyclic Puromycin Analog

113

Journal ofMedicina1 Chemistry, 1972, Vol. 15, No. 2

amine has been found to hinder acetal hydrolysis.'* The neighboring OH would not be expected to affect the rate of hydrolysis of the ketal significantly." It thus seems likely that the protonated purine moiety accounts for the difficulty of hydrolysis of compounds such as 10. When 10 was refluxed in HC1 at pH 1, black tar formed on neutralization. Similarly, treatment with Amberlite IR-120 resin at 60" for 2 hr led only to decomposition. As it a p peared that conditions necessary for hydrolysis of 10 effected decomposition of the product(s), oxime 11 was synthesized directly from 10 under what appear to be unique conditions for oxime formation. An aq soln of 10 and a 5-fold excess of HONH2 -HC1 adjusted to pH 1 with HCl was warmed (75") for 3 hr. On neutralization of the colorless reaction solution, pure 11 precipitated immediately in almost quantitative yield (solid forms as pH 3.5 is reached). However, when oxime formation was attempted by warming a solution of 10 at pH 1, and then adding HONH2.HCl along with sufficient NaOH to give a pH of 6, partial recovery of 10 (69%) and considerable darkening of the reaction solution resulted. Even in the presence of HONH2*HCl, dark tar and a lowered yield of oxime (43%) were obtained if a pH slightly lower than 1 was used. Identical conditions at a pH of 3 gave only recovered 10. These results suggest that hydrolysis of the ketal of 10 requires a pH lower than 3, but that below pH 1, 10 or its hydrolysis products is unstable and decomposes rapidly before hydroxylamine may attack. In a rather narrow pH range, HONH2 appears to prevent decomposition, probably by direct attack on the resonance-stabilized carbonium ion formed on ring .opening of the protonated ketal. Tlc of 11 indicated it t o be a mixture of syn and anti oximes. Since attempted separation by chromatography or recrystallization resulted in considerable losses, the mixture was acetylated without separation, giving a 70% yield of a chromatographically homogeneous diacetyl derivative (12). The nmr spectra of 11 and 12 confirm the structures shown. In particular, the appearance of H-2 as a doublet (J2,3= 9.0 Hz for 11 and 9.5 Hz for 12) considerably downfield from the H-3 multiplet# indicates that the oximation conditions did not result in isomerization via an enediol. The 0-acetyl oxime 12 was reduced to a mixture of amino alcohols with diborane in THF by a modification of the procedure used by Feuer and Braunstein for the conversion of cyclohexanone 0-acetyloxime to cyclohexylamine.20 This mixture was acetylated in Ac20 and the resulting acetamides, 13 (45%) and 14 (4%), were separated by chromatography and characterized. The AcNH and OH groups were shown to be cis in 13 and trans in 14 by acyl migration studies (see Experimental Section). The diacetyl derivatives 13a (39%) and 14a (3%) were isolated when the mixture of amino alcohols from reduction of 12 was treated with Ac20-pyridine, and 13a was converted to 13 on treatment with NH3 in MeOH. Hydrolysis of 13 with Ba(OH)2 gave a carbocyclic analog of the puromycin aminonucleoside, characterized as its AcOH salt 16. The cis stereochemistry of the H2N and OH groups was further confirmed by facile formation of the cyclic carbamate 18 on treatment of the carbobenzoxy derivative 17 with NaOMe in DMF. The diborane reduction of 10 appears to be the first reported example of the reduction of an oxime to an amine in the presence of a purine ring. Predominant attack of ~~~

~~~

#In the nmr (DMSO-d,)of trans-2-acetoxy-3-(6-dimethylamino-9puriny1)cyclohexanone 0-acetyloxime, the cyclohexyl analog of 12, H-2also appears as a doublet ( 7 , 3.77, .I?,B= 10.5 Hz) downfield from the H-3 multiplet at 7 5.1; unpublished results.

hydride from the purine side of the cyclopentyl ring, if general, could be useful in the synthesis of amino sugar nucleosides. It is hoped that further studies now in progress on the diborane reduction of analogs of 10 will provide information on the mechanism of this highly stereospecific reduction.** The carbocyclic aminonucleoside analog 16 was coupled by to N-benzyloxycarbonyl-pmethoxyphenyl-~-alanine~ 2 methods: A, the dicyclohexylcarbodiimide-N-hydroxysuccinimide method21'22and B, a modification of the mixed anhydride method suggested by Anderson, et a123The resulting carbobenzoxy blocked diastereomers 2a and 19a (97% by method A, 77% by method B) could not be separated. The mixture of amino alcohols from the reduction of 12 was also coupled to N-benzyloxycarbonyl-pmethoxyphenylL-alanine by method A, giving a yield of 2a and 19a, after chromatography, comparable with the overall yield via 13 and 16. Following hydrogenolysis of the Cbz group, separation of diastereomers 2 and 19 by chromatography was possible. Structure 2 is assigned to the diastereomer having [aI2'D of -83" and structure 19 to the diastereomer having [(uI2'D of -8" (see Results and Discussion). The two coupling methods resulted in samples of 2 and 19 with identical optical purities. The mass spectra of 2 and 19 are almost identical, differing slightly only in the relative intensities of some ions. The molecular ion (m/e 439) is relatively small, and there is a minor (M - 18) peak due to loss of H20 involving the OH group. As with puromycin,24the fragmentation is dominated by the aminoacyl moiety. Fission of the bond a to the C=O

B

B

HNH OH

HN OH

c=o+ I/

e=

I

I

.

I

+

m/e 289 (M - 150)

m/e 439 (MI)

Hi

B

B

OH

HN OH

I

I

B

I

HN I

?=O .+

c=o

c=o

CH-NH,

CH

CH

I

I

u

m/e 439 (M,)

m/e 318 (M - 121)

m/e 300 (M - 121 - 18)

in the molecular ion M Iaccounts for the m/e 289 (M - 150) peak. Cleavage of the benzylic bond followed by loss of H20 accounts for the base peak at m/e 300 - 121 - 18) and prominent peaks at m/e 318 (M - 121) and m/e 121. There is a metastable peak for the transition 3 18' + 300+t 18. The (M - 121 - 18) ion is somewhat less abundant for puromycin. This would be expected as the puromycin (M - 121

__

_______--

**Further details concerning this reaction will be published.

114 Journal of Medicinal Chemistry, 1972, Vol. 15, No. 2

Daluge and Vince

c

mle 228

~

/

0 4

r

-

0 06lnM 0-

mle 245 mle 271

18) ion has several favorable routes for further fragmentation which these carbocyclic analogs lack, e.g., loss of the elements of CHzO involving the 5’-OH. Fission of the m/e 300 ion on either side of the amide carbonyl accompanied by proton transfers accounts for the m/e 245 and 271 peaks. Fission of the bond between the aminoacyl group and the cyclopentyl ring accounts for the m/e 228 ion. There is a metastable peak for the transition 300’ + 228’ t 72. As with puromycin, the m/e 300 peak further fragments to give a prominent m/e 164 (B t 2H) peak characteristic of the dimethylaminopurine moiety, as indicated by an appropriate metastable peak. Other fragmentations directed by the purine moiety account for ions of m/e 206 (B t 44), 190 (B t 28), 163 (B t H), 162 (B), and 134 (B t H - CH3N). These ions, or corresponding ones, are prominent in the spectra of adenosines” and are also abundant in the spectrum of puromycin. It is consistent with the structure proposed for the (B t 30) ion in adenosines, which incorporates the ribose ether that replacement of 0 by CH2 results in a shift of 2 mass units to (B t 28). -

Results and Discussion

Antimicrobial testing of diastereomers 2 and 19 revealed that 1 isomer was completely inactive while the other exhibited growth inhibition on the same order of magnitude as puromycin. The absolute stereochemistry of the active isomer has tentatively been assigned that of structure 2 on the basis of its biological activity and in accordance with the stereochemistry of puromycin. The minimum inhibitory concns by a 2-fold serial dilution test in broth for puromycin and the carbocyclic analog 2, respectively, are as follows (mM):Staphylococcus aureus (NRRL B-3 13), 0.244 and 0.244; Bacillus subtilis (NRFX B-549, 0.030 and 0.060; Klebsiella pneumoniae (NRRL B-l17), 0.485 and 0.485; Escherichia coli (NRRL B-2 10) 0.060 and 0.120. A growth curve for S. aureus in the presence of different concns of puromycin or the carbocyclic compound is illustrated in Figure 1. A lag period is observed with both compounds when the concns are lower than those required for complete inhibition. Such a lag period is consistent with the mechanism of action of puromycin2 since the antibiotic would be expected to be consumed as it is incorporated into the growing peptide chains. The aminonucleoside 16 was tested for nephrotoxicity in rats at a dose of 33 mg/kg under the same conditions that are required for puromycin aminonucleoside to cause severe nephrotic syndrome at 15 mg/kg.26 No nephrotoxicity was observed even after 17 days of treatment with 16. The novel puromycin analog 2 provides a molecule with the structural features required for puromycin-like antimicrobial activity. Thus, the ribofuranosyl ring can be replaced with the more hydrolytically stable cyclopentane ring without loss of activity. In addition, the removal of the CHzOH moiety is not detrimental to activity and at the same time provides 2 with a resistance to kinase activity upon re-

,g_ ; - :-& - t - -:- - - - - &

-

0 24dnU

lease of the aminonucleoside. This resistance to phosphorylation circumvents the nephrotic syndrome associated with puromycin aminonucleoside. This carbocyclic analog and others which are under preparation are being evaluated for in vitro inhibition of protein biosynthesis in an attempt to explore the requirements for binding to ribosomes. Preliminary studies with 2 and 19 on an E. coli ribosomal system are consistent with the antimicrobial activities. The details of these experiments will be the subject of a future paper. Experimental Section?? 6-Oxabicyclo[3.1.0]hexan-2-one Ethylene Ketal ( 5 ) . The method of prepn of 4 is a modification of that of Guss and Rosenthal.’, Cyclopent-2-enone ethylene ketal (3)’, (37.85 g, 0.300 mole), NBS (53.40 g, 0.300 mole), NaHCO, (4.20 g, 50.0 mmoles), Et,O (240 ml), and H,O (240 ml) were stirred vigorously for 6.5 hr, at which time all of the solid had disappeared and the pH was approx 7. The aq layer was satd with NaCl, and the Et,O layer then sepd. The aq layer was extd with addnl Et,O (2 X 100 ml). The combined Et,O layers were washed with satd NaCl and dried (CaSO,). Evapn left crude bromohydrin 4 as a yellow oil (70 g). In another run, this oil was partially solidified from petr ether (bp 30-60°), giving gummy white crystals: mp 42-47”; ir (Nujol) 3450 (OH), 1200-1040 cm-’ (CO). This solid could not be recrystd and darkened on standing.* Best overall yields of epoxide were obtd by using the crude oil immediately. Crude 4 was dissolved in PhH (600 ml) and refluxed with powd NaOH (36 g) for 1.0 hr. The mixt was filtered, and the black solid washed with addnl PhH (200 ml). Evapn of the combined PhH soln and wash left a pale yellow liq which was distd, giving a forerun whiqh ir showed to be a mixt of 3 and unketalized material (1.36 g),

_-

-~

??Melting points were detd on a Mel-Temp apparatus and are corrected. Optical rotations were measured at ambient temps with a Perkin-Elmer 141 automatic polarimeter; nmr, with a Varian A-60D spectrometer; ir, with a Perkin-Elmer 237B spectrophotometer; uv, with a Cary 14 recording spectrophotometer; low-resolution, 50-eV mass spectra, with a Hitachi Perkin-Elmer RMU-6D mass spectrometer (ion source temperature 250°,accelerating potential ISOOV), equipped with a direct inlet probe. TICwas run on silica gel (Eastman chromagram sheets with fluorescent indicator) in these solvent systems: A, 2% MeOH-CHCI,; B, 5 % MeOH-CHCl,; C, 10%MeOHCHCI,; D, 1 5 % MeOH-CHCI,; E, 20% MeOH-CHCI,. Prep tlc was done on 20 X Z O cm glass plates coated with 2 mm of silica gel F 254 (E. Merck, Darmstadt) and column chromatog on silica gel (Baker, AR, 60-200 mesh). Evapns were carried out in vacuo with a bath temp of less than 45’ unless otherwise noted. Solid samples were dried in vacuo (